8:00 – 5:00 VISIBILITY

• Now that I’ve got rJava/RMI working, I’m going to try building a few composite commands and parsing the results.
• And now I’m wondering if rServe makes more sense. It appears to create a separate workspace for each connection. That would work much better for a webapp. The only thing to watch out for is that the server goes down now and then, so it will have to be restarted under control of the Remote Object.
• Spent the afternoon with the customer drawing diagrams of how our data will be packaged and migrated between platforms.

7:30 -5:30 VISIBILITY

• Server Meeting today at 11:15. Don’t see how I can make that
• Over at site starting at 1:30
• Setting up rJava: http://www.mumuen.com/2010/10/running-r-in-java-with-jrirjava.html
• Got the rtest code in the rJava examples directory compiling. It turns out that the CRAN directory has an old version. Sheesh. Updated the maven repo with the version 9 jars. Added R_HOME to the path. Restarting all to register changes… No joy.
• Following the directions here: http://mithil-tech.blogspot.com/2009/11/r-and-java-jri-via-eclipse.html
  • That worked!
• Next will be to get R to run in Tomcat: Here’s where I’ll start: http://wiki.cvrgrid.org/index.php/RJava/jri_Setup
• Useful r packages so far (google “CRAN xx’ for detailed info on the package):
  • sqldf – allows SQL manipulation of tables
  • reshape – for manipulation of data types
  • gwidgets – support for interactive graphics
  • HH
  • rJava – java interaction
  • RServe – R as a webserver

9:00 – 6:00VISIBILITY

• Had a talk with John W last Friday about the increase in hours charged 4 hrs/week to 10 hrs week for catwalk. Told him that I basically couldn’t do the level of support that I had been doing for 45min/day and that I wouldn’t be coming by unless he tells me that there is a problem.
• Going to try to talk to R through Java today.
• Found a cool audio synthsizer library for the next hand – https://ccrma.stanford.edu/software/stk/
• So that I can find this the next time: Webstart, Servlet, IDE, findme!
• Interviews!
• Got RServe installed. Going to take a look at interfacing that way.

7:30 – 4:00 VISIBILITY

• Meeting with Trish Christie and Leia(?).
  • add “… for the monthly status report deliverable which is due (month) 10th”  to the generated email.
  • Add the ability to sort and filter”Create New Request” panel and “Saved Requests” (project, status, personnel roles. etc)
  • Time that a project should be included in a report request is from the first budget year to the last budget year plus the appropriation period
  • For creating (or adding to?) a request, it may be necessary to look at projects that will become active up to some period in the future, so that as projects become active, they will be included in requests
• Went over meeting action items with Dong.
• Based on a conversation with April, I’m looking at developing a Flex GUI interface to R, running on the server and accessed through rJava.
• Talked to Brian about moving my client schedule from morning to afternoon

7:30 – 5:30 VISIBILITY

• Working on setting up a meeting with Christie and Trish to go over Dong’s proposal
• Showed Dong how to set up the Web Application with server in Eclipse. He thinks that he might be able to make it work with Maven
• Endless ANOVA
• Worked a bit on the hand with Tom. Everything seems to be working. Yipee!

7:30 – 4:30 VISIBILITY

• Deployed Dong’s code. All works well, and we *can* save out excel from Flex!
• Filed out as much as I could of Brian’s SSP
• Downloaded the Apache Commons Math library source for conversion to Actionscript
• Tried the easy way and found a Java to Actionscript converter (J2AS3). It needs some wotk for handling the FastMath class. Before I go in and change it by hand, I may try to improve the parser.
• More ANOVA

7:30 – 5:30 VISIBILITY

• Found a problem with Flash Player dealing with debug statements. It’s covered here, but the gist of it is that the non-debug version of Flash Player seems to choke on debug statements – at least Error.getStackTrace() (not sure about trace().
• Adding Randomized ANOVA to stats package. It turns out that CMM is trying to have a measure of confidence in their presentation of data but they don’t have a good statistical basis for making the call. How convenient 🙂
• Testing Dong’s new PPM and hopefully burning a disk.
• Had a grand time figuring out what’s up with flash debuggering.
• Added the following to the FGMUtils pom file to get our compiler flags in. It looks like this:

<plugin>
<groupId>org.sonatype.flexmojos</groupId>
<artifactId>flexmojos-maven-plugin</artifactId>
<version>3.8</version>
<extensions>true</extensions>
<dependencies>
<dependency>
<groupId>com.adobe.flex</groupId>
<artifactId>compiler</artifactId>
<version>4.0.0.14159</version>
<scope>compile</scope>
<type>pom</type>
</dependency>
</dependencies>
<configuration>
<definesDeclaration>
<!–  this is the command line: define=CONFIG::debugging,true -define=CONFIG::release,false –>
<property>
<name>CONFIG::debugging</name>
<value>true</value>
</property>
<property>
<name>CONFIG::release</name>
<value>false</value>
</property>
</definesDeclaration>

</configuration>
</plugin>

7:30 – 3:00 VISIBILITY

• It looks like the customer uses Flash 10. Not sure about the exact version.
• Interview with Phillip Tomlin.
• Sank into a Friday stupor and was unable to get anything substantial working.

7:30 – 5:30 VISIBILITY

• Downloaded Dong’s update on my home machine, but my CD drive just died, and I can’t burn the thing.
• Good meeting with Trish and Lea(?). The bug fixes appear to be working. We spend some time on small features, and how the email generation should work. Dong has my notes right now, but I’ll add detail later.
• Meeting with Anne and the CMM folks. Went very well. We’re going to get together again to discuss in more dtail. The possibility appears to be that CMM would like to incorporate VISIBILITY, possibly by sending a RESTful request to the server containing the table to display. We, on the other hand would like to run a set of queries against the CMM data to have more data for VISIBILITY reports. I think we can have both, which might maker everyone very happy.
• Generalizing the data input, and adding another test to the statistics package. As an aside, CMM is attempting to develop confidence intervals on the data they provide to their users. Hmmm.

7:30 – 6:30 VISIBILITY

• Installed new version of PPM. Everything seems to be working, though the meeting to walk through was postponed until tomorrow. The Combobutton is still in
• Worked on actionscript statistics classes.
  • Completed the Randomized Test
  • Updated the data sources to take user data as well as generating random

7:30 – 5:30 VISIBILITY

• Mike Cowan leaves S2F this Friday.
• Got a few more screenshots from Trish and Christie. Fortunately they have nothing new in them. Dong is still on track for a disk by COB
• Working on getting RemoteObjectTestbed2 working. This one has no java code in it and has logging. Got both working. Next, projects.
• Taking a detour into building a few statistics actionscript classes.

7:30 – 5:00 VISIBILITY

• Meeting with Trish, Christie, and someone new who I can’t remember right now.
  • Bugs – several, all hopefully having to deal with DB configuration. I gave Dong the screenshots
    • Also, add two more lines under mailing address in funding request
    • Funding request warning dialog has “financial” misspelled as “fianancial”.
    • Add “RTAs” to funding type under “detailed Information” tab
    • Selecting names seems to offer up only 20 or so names. What is the selection criteria? The same list appears for all selection buttons
    • Add check so that no start date be saved that is after an end date
    • The “Add Funding” dialog needs a sortable list for contract number – it’s a very long list. It’s also too narrow – there are some very long contract numbers. Panel is too small as well.
  • Next highest priority is the “email generator” This will produce messages to copy and paste into messages to the Financial and technical POCs
    • Create an email like the one in Funding Request for each POC that has the current financial data that they have to fill in, and the associated forms if they are not on goal
    • Have a list of all projects that the user can move through to create all emails in and easy way
    • List should have a dropdown that lets the user select the year and reporting period. Selecting a time period (FY/Quarter?) will bring up the list of POCs and the status
    • Users should be able to indicate the state of the contact:
      • Nothing done
      • Email sent out
      • Response back
      • Response entered – incomplete (if a goal has not been met and there is no justification, this should be flagged automatically. There may be an additional button if this is the case that asks the POCs for justifications
      • Response entered – complete
  • Status panel notes:
    • PPM calculated outlay should be a “checksum” of numbers. The true PM Actual Outlay needs to be entered by hand
    • Available balance is the percentage of the “last remaining to distribute” divided by the total budget
    • Goals are calculated as the total budget times the “out year” in the appropriation
• Worked on migrating DbObject classes out of the sandbox and into the utils jars/swcs

12:00 – 6:00VISIBILITY

• GREs this morning
• Working on InfoObjectManager.as
• Got an overview of the new PPM capability from Dong so that I can talk intelligently to Christie and Trish on Monday
• Interview at 4:00
• More InfoObjectManager.as. I think I have the actionscript side working.

7:30 – 4:30VISIBILITY

• Deployed the new PPM and while testing it, found that there was some odd behavior in the comments. Took a screenshot and gave it to Dong, who was also perplexed.
• Adding table updates (adding and deleting columns) and drop table capability to infoObjectManager.java – done
• Need to add the same capability to InfoObjectManager.as.
• It looks like we might need to write a tool that lets folks create update files that can then be emailed back to the financial folks for upload into the PPM database. Going to a meeting to discuss same
• Added generic query to infoObjectManager.java.
• Added check for “NULL” values when doing conversions from objects to sqlInfoObjects